Enterprise MCP & AI Agent Function-Calling Policy Benchmark v1 (3,000-Row CSV)
Updated Nov 16, 2026
Source, license and coverage
Supplier documentation. These claims are separate from the automated sample score. A listing edit date is not a data freshness date.
- License
- Oxford Digital Commercial Data License v1.0
- Source / creator
- Oxford Digital
- Collection method
- Programmatically synthesized and strictly JSON-validated MCP JSON-RPC 2.0 tool schemas, multi-tenant OAuth session contexts, and Pydantic v2 / Zod deterministic policy validators; SHA-256 deduplicated.
- Coverage start
- Not applicable
- Coverage end
- Not applicable
- Data last updated
- Not documented
- Update schedule
- Fixed v1 commercial benchmark release (3,000 verified records)
Sample structure score: 100 / 100
This automated check describes the inspected sample, not factual accuracy, legal rights, representativeness, or the quality of the entire dataset. It is not a customer rating.
Assessed 10 sample records (CSV) on 2026-11-16. All records in the provided sample were checked.
| Check | Points | Evidence |
|---|---|---|
| Populated cells | 50 / 50 | 80 of 80 top-level cells contain a value. Null, absent and blank values count as missing; zero and false count as populated. |
| Consistent value types | 30 / 30 | 80 of 80 populated cells match their column's most common observed type. Types are inferred, not checked against real-world truth. |
| Consistent record shape | 20 / 20 | 10 of 10 records have the expected fields. CSV/TSV use the header width; JSON uses the union of observed keys. |
Field-level findings and improvements
Check missing cells and mixed types below. Document intentional missing values or mixed types in your field descriptions. Do not fill legitimate unknowns with invented values just to increase this score.
| Field | Missing cells | Most common type | Other populated types |
|---|---|---|---|
| record_id | 0 / 10 | string | 0 / 10 |
| domain | 0 / 10 | string | 0 / 10 |
| policy_outcome | 0 / 10 | string | 0 / 10 |
| mcp_tool_schema_json | 0 / 10 | string | 0 / 10 |
| caller_session_context_json | 0 / 10 | string | 0 / 10 |
| candidate_tool_call_json | 0 / 10 | string | 0 / 10 |
| policy_rationale | 0 / 10 | string | 0 / 10 |
| validator_guardrail_code | 0 / 10 | string | 0 / 10 |
About this data
Curated by Oxford Digital (oxforddigitals.com). 3,000 schema-validated JSON-RPC 2.0 Model Context Protocol (MCP) and AI Agent tool-call policy evaluation & fine-tuning records across 8 enterprise verticals (FinTech & Billing, Cloud IAM, DevOps CI/CD, Healthcare FHIR, Enterprise CRM, SQL Analytics, Workspace Calendar/Mail, and Kubernetes Ops). Balanced across 5 ground-truth policy outcomes (600 rows each: ALLOW_AUTO, REQUIRE_HITL_APPROVAL, DENY_SCOPE_MISMATCH, DENY_TENANT_ISOLATION, DENY_QUOTA_OR_BUDGET_CAP) with paired Python Pydantic v2 and TypeScript Zod deterministic guardrail validators for SFT, DPO, GRPO, and policy-router Precision/Recall benchmarking.
Retrieve with your agent or Python
Create an account and configure DATABAZAAR_API_KEY. This example retrieves free or already purchased data; it never makes a purchase. For a multi-file dataset, choose a file index from the manifest.
Download the Python examplepython3 retrieve-dataset.py f10c780b-f67d-4ff4-88d8-66af0403c7b5 --output dataset.bin
Schema
| Name | Type | Description |
|---|---|---|
| record_id | string | Deterministic SHA-256 record identifier (oxd_mcp_...) |
| domain | string | Enterprise vertical (FinTech, Cloud IAM, DevOps, Healthcare FHIR, CRM, SQL Analytics, Workspace, Kubernetes) |
| policy_outcome | string | Ground-truth policy decision (ALLOW_AUTO, REQUIRE_HITL_APPROVAL, DENY_SCOPE_MISMATCH, DENY_TENANT_ISOLATION, DENY_QUOTA_OR_BUDGET_CAP) |
| mcp_tool_schema_json | string | Validated MCP tool JSON Schema and required OAuth scope |
| caller_session_context_json | string | Authenticated session context (tenant_id, role, granted scopes, HITL limit, hard quota cap) |
| candidate_tool_call_json | string | JSON-RPC 2.0 tools/call payload |
| policy_rationale | string | Architectural policy decision explanation |
| validator_guardrail_code | string | Production Python Pydantic v2 / TypeScript Zod deterministic guardrail validator |
Sample Data
Download a free sample to inspect the data before purchasing.
Public sample only. Sign in to retrieve the full dataset, including free datasets.
For AI Agents
# 1. Add to your agent's MCP config (claude_desktop_config.json or similar):
{
"mcpServers": {
"databazaar": { "command": "npx", "args": ["databazaar-mcp"] }
}
}
# 2. Your agent can then call:
search_datasets({ query: "Enterprise MCP & AI Agent Func" })
// Found: f10c780b-f67d-4ff4-88d8-66af0403c7b5
buy_now({ dataset_id: "f10c780b-f67d-4ff4-88d8-66af0403c7b5" }) // uses your saved card# Authenticate with your API key:
curl https://api.databazaar.io/datasets/f10c780b-f67d-4ff4-88d8-66af0403c7b5 \
-H "Authorization: Bearer YOUR_API_KEY"
# Purchase:
curl -X POST https://api.databazaar.io/datasets/f10c780b-f67d-4ff4-88d8-66af0403c7b5/buy \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"dataset_id": "f10c780b-f67d-4ff4-88d8-66af0403c7b5"}'